How can We Use Mechanism-based Receptor Models for Designing First in Human Studies?

First in man studies, are often considered as intended primarily to explore tolerability and pharmacokinetics. These are important objectives but there is often much that can also be done in first in man studies to explore the relationships between dose, exposure and effect(s) and to identify doses with the potential to be clinically effective in addition to well tolerated.
Mechanism based modeling already plays a major role in designing first in human and other early development studies. Physiologically based pharmacokinetics (PBPK) modeling can predict probable human PK and wherever possible it should be used to justify the dose(s) and sampling schedule for first in man studies. The concepts can be applied to many small molecules and also some antibodies, especially for those expected to show target mediated drug disposition in humans. However predicting drug effects requires additional models to link exposure and effect. This may be through combining PBPK and PKPD models developed from non-clinical studies, with appropriate scaling to account for between species differences or the effects of disease. In some cases it is possible and useful to develop more complex, receptor-based mechanistic models and this is likely to increase in the future. This model development will be fueled by advances in "-omics" technologies and wider availability of patient data, both from intensively characterised disease cohorts and from internet-based, crowd-sourced, patient-driven databases. Such models will provide more support for the decision to progress to clinical trials; better predict dose/exposure/response curves to use as informative priors for adaptive dose-finding study designs; help determine the required level of pharmacological activity and effects on biomarkers for dose(s) to be progressed into efficacy trials; explore and understand response variability and improve decision making in early development studies.
A major challenge with mechanistic models is incomplete understanding of human physiology, biochemistry and pathology of systems and diseases, hence the models are incomplete too. For poorly understood diseases and drug mechanisms of action this limits the extent to which mechanistic models are accepted for decision making. New data will improve the predictive value of the models but uncertainties will remain. However, this should not be used an excuse not to use models. Instead, the design of the first in man studies should obtain data to allow the models to be refined and improved in order to increase confidence in their use for subsequent PoC studies.